A systolic array architecture consists of a grid of simple processing elements (PE) connected through local interconnects. With a massive number of PEs and a local interconnection, such an architecture is capable of achieving high performance and energy efficiency. My dissertation focuses on extending the research of systolic array architecture in two fields: automatic systolic array synthesis and architecture optimization. The first part of the dissertation focuses on the automated systolic array synthesis. Designing high-performance systolic arrays requires an understanding of both the application characteristics and hardware architecture, requiring non-trivial efforts to reap its benefits. There exists a large body of past works on developing compilation frameworks for systolic arrays. However, these works failed to reach a balance between the generality, performance, and productivity, making them hard to use in practice. Our work advances this field by leveraging two compilation technologies, the polyhedral model and high-level synthesis (HLS). We propose a new compilation framework, AutoSA, which is built on the polyhedral framework and is capable of generating high-performance systolic arrays on FPGA in HLS languages. We show that AutoSA can handle applications with complex dependence structures and generate designs with performance comparable to or better than manual designs. AutoSA incorporates a broad set of hardware optimization techniques that open up a vast design space which is intractable to explore manually. To cope with this challenge, we propose an efficient auto-tuning framework, Odyssey, which finds optimal designs within seconds. With both AutoSA and Odyssey, we reduce the development cycles of systolic arrays from weeks to days, which significantly boosts the productivity compared to the prior works. In the second part of the dissertation, we present two application optimization studies that deploy systolic arrays for various applications and platforms. The first study investigates the architecture trade-offs when using systolic arrays for one important application, convolutional neural network (CNN). The results show that a single monolithic systolic array is insufficient to handle the divergent characteristics of different CNN layers. Therefore, we further explore the use of a multi-array architecture that implements several smaller systolic arrays with different configurations customized for each CNN layer. Multi-array systems help improve the throughput with a cost of longer latency. This work reveals the complexities and trade-offs when mapping a real-world application to systolic arrays. In addition to FPGA, systolic arrays can also be mapped to GPU as an overlay above the existing GPU architecture. The second work investigates the performance trade-offs when mapping systolic arrays to GPU. We achieve a performance speedup by leveraging the shuffle instructions on Nvidia GPUs to implement the inter-PE communication compared to baselines using the shared memory. Systolic array architecture plays an important role in the post-Moore's law era as one architecture candidate capable of delivering high performance and energy efficiency. The works presented in this dissertation provide comprehensive and efficient solutions to lowering the programming efforts and optimizing the performance of this architecture. We hope the promising results from these works will open the door to more deployment cases of systolic arrays in a broader range of applications and hardware platforms in the future.

As collaborative governance is initiated in different countries and across diverse policy fields, researchers are examining how the formation, operation, and outcomes of collaborative governance are conditioned by the context and time in which it takes place. Scholarship increasingly tests collaborative governance frameworks and examines their generalizability in different social, economic, political, cultural, and environmental contexts. Some scholars are taking a historical approach to collaborative governance research and theory by emphasizing the role of time in shaping collaboration. Changes in both internal conditions and external circumstances can drive the metamorphosis of elements within collaboration.

My dissertation follows a three-paper approach, with a series of theoretical and empirical chapters that explore questions related to context and time in collaborative governance, by drawing on the specific case of China. Specifically, I highlight the motivation, adoption, and evolution of collaboration and investigate the following three research questions: (1) What motivates actors to collaborate in authoritarian China? (2) What factors can drive governments to adopt collaborative arrangements? (3) How do collaborative actions evolve in self-initiated versus mandated collaborations?

Chapter 2 highlights the role of political context in shaping motivation of actors from the public sector, the private sector, and civil society to work together collaboratively. Most studies on motivations for collaborative governance come from Western, democratic contexts, and a comprehensive assessment of motivations in authoritarian contexts is missing. I bring political context in by investigating what motivates state and non-state actors to collaborate in China, a classic example of an authoritarian state. By conducting a systematic review of 264 empirical studies on collaborative governance in China published from 2006 to 2021, this chapter reveals prevalent motivators including vertical commands, material-resource dependence, rules and regulations, legitimacy, economic benefits, and political resources; and uncommon motivators including asset specificity and share beliefs for both state and non-state actors in Chinese cases.

Chapter 3 uses the case of River Chief System (RCS) in China to investigate the determinants behind governments to adopt a collaborative arrangement initiated in a campaign-style manner. I incorporate a multi-level framework, including top-down effects, bottom-up effects, and horizontal effects, as well as internal determinants, to understand the motivations behind adoption by two distinct types of subnational governments (provincial and prefectural governments). Using event history analysis, this research examines why subnational governments adopted the RCS voluntarily between 2006 and 2016. Results confirm the significance of multi-level factors, such as density and time duration of local adopters, provincial environmental investments, and leaders’ previous experience with innovation, in shaping collaborative policy adoption, and also reveal different incentives behind provincial and local governments’ policy innovations.

Chapter 4 uses the case of RCS to investigate whether and how mandated versus self-initiated collaborations differ in their adoption of collaborative actions over time. I analyzed the temporal distribution of collaborative actions in 95 prefectural cities over fourteen years, employing panel data analysis to predict factors that shape collaborative actions. Self-initiated collaborations did not significantly outperform mandated ones in taking more collaborative actions as they mature. However, receiving a provincial mandate made self-initiated cities take more collaborative actions as local water pollution increases. These findings confirm self-initiation as fostering collaborative efforts in the long term and highlight emerging drivers during the collaborative process.